Abstract
Machine translation has witnessed many advances in recent years and has become ever more accessible. However, no matter how good an automated translation is, a proficient human translator does a better job. In this project, we attempt to identify areas of machine translation that need addressing by analyzing texts from three different disciplines translated using 'Google translate' from English into Arabic. We then suggest where and when human intervention is necessary as well as further automated functions to be incorporated in an improved proposed model to be developed based on the findings of this study. Human intervention can be required in instances when machine translation fails to detect the context, and automated functions include determining the discipline of the text using keywords and titles and connecting the text to more specialized dictionaries.